The content discusses the implementation and integration of the ML-KEM post-quantum Key Encapsulation Mechanism (KEM) into the TLS 1.3 protocol. The key highlights are:
The authors present the first optimized implementation of ML-KEM using the AVX-512 instruction set, achieving a 1.64x speedup compared to the state-of-the-art AVX2 implementation. This is done by parallelizing polynomial multiplication, modular reduction, and other computationally intensive modules within ML-KEM.
The authors propose a batch key generation method for ML-KEM that can batch 8 independent key pairs, resulting in a 3.5x to 4.9x speedup compared to key generation without batching.
The authors revisit two IND-1-CCA KEM constructions (TCH and TRH) and implement them using the underlying CPA-secure PKE of ML-KEM. They evaluate the performance of these IND-1-CCA KEMs and integrate the better-performing one into TLS 1.3, demonstrating improved handshake efficiency compared to IND-CCA KEMs.
The authors integrate the optimized AVX-512 implementation of ML-KEM into TLS 1.3 and evaluate its impact on TLS 1.3 handshake performance under both hybrid and PQ-only modes. The results show that the efficient ML-KEM implementation can yield a higher number of handshakes per second compared to the latest AVX2 implementation.
Naar een andere taal
vanuit de broninhoud
arxiv.org
Belangrijkste Inzichten Gedestilleerd Uit
by Jieyu Zheng,... om arxiv.org 04-23-2024
https://arxiv.org/pdf/2404.13544.pdfDiepere vragen